The unemployment rate improved last month because 264,000 people stopped being counted.
I once spent eleven weeks trying to fill one warehouse lead role at a consumer products company I helped build. Four applicants, two no-shows, and a lot of me insisting the labor market was just tight. The real story sat in a statistic I never once checked: men not in the labor force.
The July 2026 jobs report gave that excuse a proper burial. Payrolls fell by 23,000, yet the unemployment rate still improved to 4.1% from 4.2%. That looks like good news until you read the next line, where the civilian labor force shrank by 264,000 people. The rate got better because the denominator got smaller.
Underneath that monthly noise sits a number that barely moves and matters far more. Roughly 6.9 million American men between 25 and 54 are neither working nor looking for work. That’s about one in nine men in the age band where work used to be near universal. None of them appear in the 4.1% everyone quotes.
Mike Rowe has been putting a version of this stat in front of large audiences, calling it the thing that worries him most. He’s directionally right, and the underlying math is worth walking through carefully, because the popular retellings have mangled it in both directions.
The Number Rowe Uses, and What BLS Actually Shows
Rowe’s figure is 7.2 million “able-bodied men” in their prime working years who aren’t employed and aren’t searching. He said it in an interview with Moms for America, and he didn’t cite a source. The order of magnitude holds up anyway.
Here’s the arithmetic from the July 2026 household survey. Prime-age men, meaning ages 25 to 54, had a labor force participation rate of 89.2% and 57.0 million people in the labor force. That implies a population of about 63.9 million. Subtract the labor force and you’re left with roughly 6.9 million men outside it entirely.
The employment side confirms it. In practice, about 54.8 million prime-age men were employed in July, an employment-population ratio of 85.8%. So 14.2% of that age group, around 9.1 million men, did no paid work at all. Of those, 2.2 million looked for a job and got counted. The other 6.9 million didn’t and weren’t.
Two words in Rowe’s framing don’t survive contact with the data. “Able-bodied” is an assumption, not a finding, since the survey never asks that question. Rowe also implies these men are unemployed, which is precisely the category they are not in. That distinction sounds pedantic until you realize it’s the entire reason the stat is invisible.
Worth flagging one small coincidence: the total number of unemployed Americans of every age and sex in July was also about 6.9 million. One of those figures leads every news cycle, and the other one shows up when a TV host brings it up.
How the Unemployment Rate Makes People Disappear
The unemployment rate answers a narrower question than almost anyone assumes. It isn’t wrong, it’s just built to measure job search, not job holding.
What Counts, and What Doesn’t
To land in the unemployed bucket, you need three things at once. You’re without a job, you’re available to work, and you actively looked in the prior four weeks, unless you’re on temporary layoff. Miss any one of those and you fall into “not in the labor force,” which BLS abbreviates as NILF.
The rate itself is unemployed divided by the labor force, where the labor force is employed plus unemployed. Notice what happens when someone gives up. They exit the numerator and the denominator simultaneously, so the rate improves. Discouragement reads as progress.
Participation and the employment-population ratio don’t have that flaw, because both start from the entire civilian noninstitutional population. Nobody vanishes from the denominator by quitting the search. I’ve made this case before, in a breakdown of what the headline jobs number misses about male employment. The participation series belongs on your dashboard right next to the rate.
July 2026 Was a Live Demonstration
You rarely get a textbook example this clean. In July, employment fell by 87,000 and the number of unemployed fell by 178,000. Both of those moved down, so the labor force contracted by 264,000.
The unemployment rate improved by a tenth of a point on that. Overall participation slid to 61.4%, its weakest reading since early 2021. If you only tracked the headline, you’d have read a shrinking workforce as a strengthening one.
What Men Not in the Labor Force Actually Say
Health dominates their answers, and it isn’t close by any measure. The best public data comes from a BPC-Artemis survey of 2,165 non-working adults ages 20 to 54. Artemis fielded it in July 2023, and the Bipartisan Policy Center published the men’s breakdown in February 2024.
Among prime-age men out of the labor force, 57% gave physical or mental health as their main reason. That splits into 55% citing a disability, serious illness, or receipt of disability benefits, plus 2% citing a mental, emotional, or behavioral reason. For comparison, only 16% of men who were unemployed and still looking said the same thing.
Skills came second, and they hit hard. Nearly half, 47%, pointed to obsolete skills, insufficient education, or a rough work history as a barrier. Among unemployed men who were still searching, that figure climbed to 67%.
Caregiving shows up more than the stereotype allows. Just over 30% of non-working prime-age men cited caring for a seriously ill, disabled, special needs, or aging family member. That share barely differed between searchers and non-searchers.
The Telephone Game Around This Survey
A version of these findings has circulated widely with one number quietly flipped. Several outlets reported that close to 30% of these men said they weren’t working “by choice,” with a separate 9% attributed to caregiving. BPC’s own writeup of the survey says the just-over-30% figure belongs to caregiving, and it never frames a “by choice” category that way.
I chased that discrepancy for twenty minutes before finding the original post, which is twenty minutes I’d like back. The lesson isn’t that anyone lied. Secondary coverage compresses survey categories, and the compressed version travels faster than the source.
One more nuance the retellings drop: the 57% is a main-reason question, while the caregiving and skills figures are cited-as-a-barrier questions. Those don’t sum to 100% and were never meant to.
The Five Numbers, and What Each One Hides
Measure | July 2026 | What It Counts | What It Quietly Leaves Out | Worth Watching? |
Headline unemployment rate | 4.1% | People without a job who searched in the prior four weeks | Everyone who stopped searching | Only alongside participation |
Prime-age male unemployment rate | About 3.9% | 2.2 million men aged 25 to 54 actively looking | The 6.9 million who are not looking | Same caveat, smaller lens |
Prime-age male participation rate | 89.2% | Employed plus searching, over the whole population | Nothing. The denominator is fixed | Yes. This is the one |
Prime-age male employment-population ratio | 85.8% | 54.8 million men aged 25 to 54 actually working | Nothing. Also a fixed denominator | Yes. Pair it with participation |
Prime-age men not in the labor force | About 6.9 million | Men 25 to 54 neither working nor searching | Reasons. The survey does not ask if they are able-bodied | Yes, as a level, not a rate |
Prime-age men not working at all | About 9.1 million (14.2%) | Unemployed plus not in the labor force | The line between cannot and will not | Yes. This is the honest total |
Why They Say They’re Out, in Their Own Categories
Reason Given | Share of Prime-Age Men Out of the Labor Force | What It Actually Means | Does It Support the Able-Bodied Framing? |
Physical or mental health (main reason) | 57% | 55 points from disability, serious illness, or disability benefits; 2 points from mental or behavioral health | No. Largest single block, and it is medical |
Obsolete skills, education, or work history | 47% | Cited as a barrier, not necessarily the main reason | Partly. These are men retraining could reach |
Caring for a family member | Just over 30% | Roughly the same rate among searchers and non-searchers | No. This is unpaid work, not idleness |
Health insurance would matter a lot for returning | 52% | Benefits, not wages, are the stated re-entry lever | Mixed. It implies a reachable group |
Mental health benefits would matter a lot | 40% | Overlaps heavily with the health block | No |
Where Nine Percentage Points Went
Nobody has fully solved this, so be suspicious of anyone who says otherwise. Participation for prime-age men has fallen from roughly 98% in September 1954 to 89.2% today. That’s a nine-point slide across seven decades and a dozen business cycles.
The candidates are familiar. Manufacturing employment shifted and disability enrollment grew. Meanwhile incarceration rates rose and then partly fell, marriage rates dropped, and the wage premium on a high school diploma eroded. Each explains a slice, and none explains the whole thing.
Where researchers split is on the labor market itself. Scott Winship at AEI has argued the decline hasn’t come from any deterioration in the economy, pointing to survey answers about health and choice. Others read the same skills numbers as evidence that demand moved and a lot of men didn’t get to move with it.
My read, for what an operator’s read is worth: the education gradient should worry you most. The decline concentrates among men without a bachelor’s degree, and 47% name obsolete skills or a thin work history as a barrier. That’s a supply problem you can actually address.
Why the Answer Matters for Hiring
Attribution drives the fix. If the cause is health, the lever is coverage, accommodation, and treatment. Skills call for training and apprenticeships instead. An incentives story would point you at program design.
Right now the survey data points hardest at health and second-hardest at skills. Meanwhile most of the public argument runs on the third theory. That gap between what the data says and what people fight about is the reason this post exists.
What the Days Actually Look Like
Time-use data is where this gets uncomfortable, so let’s be precise instead of preachy about it. Brookings analyzed pooled American Time Use Survey data for prime-age adults. Men out of the labor force with no child at home average about nine hours a day of leisure. Roughly 70% of that is screen time.
Run the multiplication. Six and a third hours a day works out to roughly 2,300 hours a year in front of a screen. That’s more than a full-time job. That’s the figure that tends to end dinner-table debates, and it’s also the one most likely to be misread as a moral verdict.
Causation runs both ways here, which the outrage version of this story never mentions. Chronic pain, depression, and disability all produce sedentary days. Sedentary days also make re-entry harder. For the broader context, the screen time breakdown by age group covers the general population. Prime-age non-working men sit at the far end of a distribution that has moved for all of us.
The honest read is that this cohort has enormous unstructured time and very little institutional pull on it. Meanwhile, employers keep designing roles as though everyone’s on a fixed 40-hour schedule in a fixed location.
What This Costs You as an Operator
Start with the hiring math, since that’s where you feel it first. Total job openings sat at 7.4 million in June 2026, and construction alone reported more than 300,000 openings, up 81,000 from a year earlier. Notably, those roles skew toward exactly the demographic that’s checked out.
That’s the paradox Rowe keeps pointing at. Nationally, hiring has cooled and payrolls are shrinking. Specific trades still can’t fill seats. So “the labor market is soft” and “I can’t find a foreman” are both true at once.
Beyond that, the second-order costs land slower and hurt longer. Lower participation shrinks potential GDP and narrows the tax base, and the federal budget already leans hard on a smaller share of earners. Households absorb the rest through a single earner carrying two people’s expenses.
There’s a practical angle too, and it’s the part I’ve actually used. The stated re-entry levers in that survey were health insurance, sick leave, disability accommodation, and schedule flexibility. Not one of them is a raise. Say you’re a small employer losing a comp war you were never going to win. Then flexible arrangements and the productivity data behind them cost you less than another dollar an hour.
The Term of Art Worth Keeping
In short, if you take one habit from all this, make it a metric swap. Track prime-age male participation and the employment-population ratio, not the unemployment rate alone.
The unemployment rate is a search statistic. Participation is a work statistic. When those two diverge, the economy is saying something the headline was never built to say. It’s the same complaint I have about how official inflation math handles your actual costs.
The Number I Watch Instead
Prime-age male participation peaked near 98% in September 1954. It sits at 89.2% today. That’s not a business cycle. In reality it’s seventy years of structural drift, and no monthly print is going to reverse it.
So here’s my honest position, for whatever it’s worth. Rowe’s 7.2 million is close enough that arguing about the decimal misses the point. His “able-bodied” adjective, though, is doing work the data won’t support. Most of these men aren’t sitting out because they’re lazy, and a meaningful share aren’t sitting out for reasons anyone would choose.
Even so, that leaves millions of men the labor market could plausibly reach. It also leaves employers who can’t fill a single opening. Both facts survive the fact-check.
If you’re hiring this quarter, pull the participation rate for your metro first. Then ask whether your benefits package answers any of the four things these men said would bring them back. It cost me eleven weeks to learn that the tight labor market I kept complaining about wasn’t really about wages. Whether you’re hiring, buying a business, or staying put, participation will tell you more about the next two years. The unemployment rate never will.
Related Posts: U.S. vs. World Population | Remote Work Productivity | The Road Not Taken